1. Field of the Invention
The present invention relates to a binary coding circuit for use in copying machines, scanners, printers, image display devices and other such equipment requiring the ability to binarize text and images clearly.
2. Description of the Prior Art
In a conventional a binary coding circuit the signal input to the circuit is passed through a low-pass filter and is then given a prescribed offset value to form a threshold signal, and binarization is carried out by comparing the threshold signal with the original signal. FIG. 3 (a) shows a conventional adaptive binarization processing waveform. In the figure the solid line indicates the input signal A, shown as a voltage corresponding to the brightness level of the signal. The dashed line represents the threshold signal B. Input signal A has a somewhat flat waveform owing to the fact that high-frequency components have been cut. When the input signal is near the black level an offset .delta. is applied to the white side, and when the input signal is near the white level the offset .delta. is applied to the black side.
Threshold signal B and input signal A are compared and if input signal A is larger than threshold signal B it is binarized to white, and if it is smaller it is binarized to black.
In this type of conventional binary coding circuit, basically background spatial frequencies are treated as low-frequency components and, as shown by FIG. 3 (a), the usual practice is to binarize the input signal by using a threshold value obtained by adding some offset value to the low-frequency components. In such cases, if the offset value is too small background noise components caused by nonuniformities in the paper surface, in the case of a document scanner, for example, will appear on the binarized image, degrading the clarity of the image. When this results in a high number of noise spots, it results in a marked degradation of the efficiency with which the image data can be compressed, thereby reducing the effectiveness of the compression. On the other hand, if the offset value is too large low contrast characters and images will not be binarized.